Multi-target prediction for dummies using two-branch neural networks

Multi-target prediction (MTP) serves as an umbrella term for machine learning tasks that concern the simultaneous of multiple target variables. Classical instantiations are multi-label classification, multivariate regression, multi-task learning, dyadic prediction, zero-shot network inference, and matrix completion. Despite significant similarities, all these domains have evolved separately into distinct research areas over last two decades. This led to development a plethora highly-engineered methods, created substantially-high entrance barrier practitioners not experts in field. In this work we present generic deep methodology can be used wide range multi-target problems. We introduce flexible multi-branch neural architecture, partially configured via questionnaire helps end users select suitable MTP problem setting their needs. Experimental results illustrate proposed manifests competitive performance compared methods from specific domains.